Abstract
ERA permits a determination of threshold in non-medicated as well as alimemazine tartrate sedated children. This determination can be made in spite of a demonstrated 10% rate of false positive or negative responses ('misses') during the session. Non-medication is preferable because during the sedation the range of uncertainty may well exceed 30 dB SL, whereas without medication the range will not exceed 30 dB SL. During light surgical anaesthesia the rate of false positive (19%) plus misses (12%) exceeded 30% and the range of uncertainty could exceed 30 dB SL. An improvement in the ERA techniques, i.e. selective averages calculated from sweeps with similar EEG-characteristics (frequency and amplitude), made the vertex response clearer and was helpful clincally with the non-medicated and sedated child, but during anaesthesia no improvement was found. The evaluation of errors in ERA are based on a new scoring system permitting separate evaluation of response reproduciblity and response patterns.